Light beam homogenizer designs used for decades are largely based upon lenticular lens arrays that date back to the 1940s and 1950s. A more recent design is from the late 1980s that comprises a pair of crossed-cylindrical lenticular lens structures coupled with a condensing lens. The 1980s design has 10 optical surfaces which represents a significant source of optical loss, long beam path (typically ~ 1.5 meters) and high cost from having so many optical elements. Further, this design requires careful and precise alignment of the light beam in relation to the optics comprising the homogenizer, in particular angular rotation of the lenslets and their positioning in the X-Y plane of the optical axis. These designs are unable to produce a homogenized ring illumination, which is useful for ablating polymer insulation from metal bonding pads in microelectronic fabrication of multichip modules and memory chips.

Figure 1 illustrates a patent pending optical device called a “Field-Mapped Beam Homogenizer” that transforms a spatially, nonuniform light beam into a spatially homogeneous profile suitable for precision laser micromachining solar concentrators, or as an illumination source for photolithography and TV projection displays.

The unique optical device “field-maps” a homogenized illumination by using spherical, cylindrical, axicon or prism optical segments — or in some applications a combination of these various elements — and placing the optical segments in a “mapped” configuration whereby light passing through each optical segment directs the light to overlap at a homogenized plane with the desired shape (rectangle, square, rectangular or circular ring illumination). The geometry of the homogenized field is limited only by the fabrication techniques used in segmenting and how the segments are physically arranged.

The field-mapped homogenizer in its basic form is fabricated from spherical or cylindrical lenses. The lenses can be either negative or positive, depending on the type of illumination, size and numerical aperture required. The lenses are segmented from larger lens elements in a predetermined way and then specific segments are selected and repositioned in a specific order so that the light passing through each lens segment recombines at the desired homogenized plane. Figure 2 illustrates this process with 3 plano-concave lenses to form a 1 × 3 segment homogenizer for simplicity of explanation.

The 1 × 3 segment homogenizer in Figure 2 would produce a 15 mm × 15 mm homogenized field at two times the absolute focal length of the negative spherical lens used, i.e., if the plano-concave lens prior to segmenting had a focal length of -100 mm, then the homogenized field would be + 200 mm away from the array after segmenting and mapping. Specifically segment “A” is extracted from the center of lens 1, which has coordinate values (-2.5 mm, -2.5 mm) and (2.5 mm, 2.5 mm) and defines the 5 mm × 5 mm square lens segment. It is assumed here that the center of each lens, prior to segmenting, would have coordinate value (0 mm, 0 mm). Next, segment “B” is extracted from lens 2 with coordinate values (-2.5mm, 0) and (2.5 mm, -5.0 mm). Lastly, segment “C” is extracted from lens 3 with coordinate values (-2.5 mm, 0) and (2.5 mm, 5.0 mm). In the assembled array segment “B” would be the top lens, segment “A” the center lens, and segment “C” the bottom lens.

The ray trace shown in Figure 2 (red, green and blue rays) illustrates the controlled refraction which forms at a common, homogeneous plane with a focal length of 200 mm. The cross-sectional view of the array (Fig. 1 ray trace) resembles a Fresnel lens. By definition each lens segment can be considered a Fresnel zone but as such are not configured in the traditional manner.

The field-mapping can be applied to prism or conical forms. Figure 3 shows an example of an axicon based field-mapped ring illumination homogenizer. The top graphic in figure 3 shows an isometric view of a segmented axicon where each section has a different conic constant that allows the light to be overlapped at a fixed distance from the array. For the given example the focus of the homogenized ring is about 120 mm and produces a 3 mm wide ring. The spot diagram at the bottom of the figure shows how the light transforms after the array. This illustrates the flexibility of the basic concept. Likewise, segmenting and mapping rectangular cylindrical lens segments, positive or negative, will produce a rectangular ring illumination (Figure 4).

A particular microelectronic application, for example, may require the illumination of the outer region of a circuit that is 13 mm × 21 mm, where metal bonding pads need to be cleared of an insulating polymer. In a typical excimer laser optical system, one might require an ablation fluence of 500 mJ/cm2 in order to expose the metal bonding pad. A field of 13 mm × 21 mm would therefore require 1.36 J of energy on target for the given fluence. This is very unrealistic for current excimer laser technology at any reasonable cost. Since the bonding pads are on the periphery of the chip, one need only illuminate the outer perimeter and not the entire die of which most of the laser’s light would be blocked by the illumination mask. Existing homogenizers would have to illuminate the entire 13 mm × 21 mm area or at lower energy, process the die in steps. This is impractical in most cases due to the added time of step and repeating the circuit with an accompanied mask change.

In order to achieve a rectangular ring illumination with field-mapping, segments from several plano-convex cylinder lenses are mapped to a homogenized field that forms a rectangular ring of 13 mm × 21 mm as an annulus. The width of the annulus could be 3 mm as exemplified in the spot diagram of figure 4. In this example the field-mapped ring illuminator would only require 465 mJ at the mask to achieve the needed 500 mJ/cm2 fluence to expose the bonding pads. This is very much achievable with current excimer technology. This clearly improves the overall efficiency of the laser’s energy over larger areas that are simply not possible with existing homogenizer designs.

The field-mapped homogenizer is fairly insensitive to minor angular errors and will have no focused pupil to damage downstream optics when employing plano-concave segments. In addition, combining a zoom telescope past the lens segments permits dynamic modification of the homogenized field. Existing designs usually can only be altered by 25 to 50% and some of the optical elements in the beam homogenizer are very much susceptible to damage from focused pupils. A 3× zoom range is easily achieved with the field-mapped homogenizer and even takes up less space in the beam delivery system. This feature further adds to the flexibility and usefulness of the design.

The field-mapping concept can be implemented on the micro level to cover larger fields for display devices whereby the lens segments are fabricated by diamond turning, electron beam or reactive ion etching or other lithographic techniques. The ability to form complex shapes and structures for more esoteric field-mapping concepts is well within the realm of possibility. Even at the macro level one could field-map large Fresnel lens segments to create even larger integrators/concentrators for solar panels. The field-mapped beam homogenizer opens up a lot of interesting illumination possibilities.

This article was written by Michael Scaggs, Executive Director, Neoteric Concepts, LLC (Weston, FL). For more information, contact Mr. Scaggs at This email address is being protected from spambots. You need JavaScript enabled to view it. or visit .